Mac Pro Tower (Nehalem or Westmere)

[nevernamed]

Hey all.

I was wondering if anyone has experience on running DIVA on Mac Pro towers released in 2012. There were two models based on two different Intel architectures: Nehalem and Westmere. One was four core with hyperthreading and the other six core.

How would these compare to a 2013 i7 iMac?

Reason I ask is I have a late 2011 iMac that is having some problems. I am faced with a choice of either addressing the problems or just cutting my losses and upgrading.

I'd like to avoid consumer grade computing platforms if at all possible so I'm thinking of maybe switching to the aforementioned Mac Pro tower. One can still buy them on the used market.

Would it be sufficient to run Diva is the question? Please chime in if you have had Diva running on one of these please.

thanks.

[aaron aardvark]

I have never heard of Mac Pro towers, so my reply may not be relevant to your question. I only have this to report with an iMac that is almost 5 years old: I tried out Mac OSX 10.9.3 and a trial version of Cubase Elements 7, and it did not work well with my Zebra, Diva, and some of my "French" synth plug-ins. I reverted back to my old Mac OSX 10.6.8 and CubaseLE4, and it is back to working much better again.

[MaF]

I do not have DIVA but on this site you should get the info you need when you want to compare the performance of Mac models: http://browser.primatelabs.com/mac-benchmarks
You can choose between 32 and 64 bit and single or multi core utilization. As you will see the single core performance of an iMac 2013 (new ones are expected soon) can be better than that of a Mac Pro (even the current ones). I guess for an instance of DIVA a faster single core performance is better than having more cores. For bigger projects, however, more cores seem to be better.

Should you go the Mac Pro route even the fastest Mac Pro from 2012 can be upgraded with faster CPUs (2nd hand from ebay) I upgraded my Mac Pro from 2009 from eight to twelve cores giving it a speed bump from 13000 to nearly 26000 points and putting it on par with the second fastest new Mac Pro in multicore performance. Not bad for a five year old computer and 300 €.

[pdxindy]

I do not have DIVA but on this site you should get the info you need when you want to compare the performance of Mac models: http://browser.primatelabs.com/mac-benchmarks
You can choose between 32 and 64 bit and single or multi core utilization. As you will see the single core performance of an iMac 2013 (new ones are expected soon) can be better than that of a Mac Pro (even the current ones). I guess for an instance of DIVA a faster single core performance is better than having more cores. For bigger projects, however, more cores seem to be better.

Should you go the Mac Pro route even the fastest Mac Pro from 2012 can be upgraded with faster CPUs (2nd hand from ebay) I upgraded my Mac Pro from 2009 from eight to twelve cores giving it a speed bump from 13000 to nearly 26000 points and putting it on par with the second fastest new Mac Pro in multicore performance. Not bad for a five year old computer and 300 €.

A quad core 2.8 Ghz Macbook Pro is faster than the current quad core 3.7 Ghz Mac Pro???

I'm doubtful that is so...

[MaF]

As I said, in single core performance several models are superior to the Mac Pro, though it seems surprising. However, it makes sense when you compare these new MacBooks to the last generation (Late 2013). Future benchmarks from other users will clarify. These MacBooks were just released.

[pdxindy]

As I said, in single core performance several models are superior to the Mac Pro, though it seems surprising. However, it makes sense when you compare these new MacBooks to the last generation (Late 2013). Future benchmarks from other users will clarify. These MacBooks were just released.

According to that chart, the Macbook Pro @ 2.8 ghz is faster than the Mac Pro @ 3.7 Ghz for both single core and multi-core...

How can that be right?

[nevernamed]

Here is a Mac Pro (tower) circa 2012:

http://www.everymac.com/systems/apple/m ... specs.html

It's a popular computer. I've had one in the past in fact. The true reason why I'm considering one is I suspect it will outlast any consumer-grade device apple has out on their shelves. Asset renewal cycles in the pro-audio world and in technology are not easily reconcilable. Needless to say I am tired of upgrading my computer every 2 years. Asset renewal cycles must be longer. I want a device that can last and be problem free for at least 5 years. I'd also like software developers to contribute to longer asset renewal cycles. Look I'm only asking for personal testimony. If you have a mac pro from 2012 and you run Diva on it I'd like to hear from you. I think it should run fine but before I commit it would be useful to hear from someone with some experience. Benchmarks are theoretical.

[MaF]

@ nevernamed: I totally agree with you on the life cycles of hard- and software and that benchmarks won't tell you how many instances of DIVA you will be able to run on the 2012 MacPro (12x2.4GHz). But they will give you a hint how it compares to an iMac (late 2013) in general since this was a question in your OP (even though DIVA related and not focused on benchmarks).

@ pdxindy: This new (fastest) Macbook is indeed faster than the slowest MacPro from 2013 in multicore performance, but they are so close - would you notice it? GHz is not everything. The different architecture of the CPUs makes a difference too. Whatever the reasons is, there's a reason.

[pdxindy]

@ pdxindy: This new (fastest) Macbook is indeed faster than the slowest MacPro from 2013 in multicore performance, but they are so close - would you notice it? GHz is not everything. The different architecture of the CPUs makes a difference too. Whatever the reasons is, there's a reason.

Or maybe that test is not right...

[MaF]

I cannot refute this notion. The only thing I know about Geekbench is that the Mac community more or less relies on it.

[jdnz]

in terms of single core performance, we ran tests of the s/w we use on our 12 core macpros at work on a current i7 imac, and the imac ran rings around the macpro (presumably down to a combination of turbo-boost on the i7, and the faster memory bus on the haswell cpu).

[chilly7]

I have new cilinder Mac Pro base model.
If u interested ask me and i can take a look for u.

[chilly7]

I do not have DIVA but on this site you should get the info you need when you want to compare the performance of Mac models: http://browser.primatelabs.com/mac-benchmarks
You can choose between 32 and 64 bit and single or multi core utilization. As you will see the single core performance of an iMac 2013 (new ones are expected soon) can be better than that of a Mac Pro (even the current ones). I guess for an instance of DIVA a faster single core performance is better than having more cores. For bigger projects, however, more cores seem to be better.

Should you go the Mac Pro route even the fastest Mac Pro from 2012 can be upgraded with faster CPUs (2nd hand from ebay) I upgraded my Mac Pro from 2009 from eight to twelve cores giving it a speed bump from 13000 to nearly 26000 points and putting it on par with the second fastest new Mac Pro in multicore performance. Not bad for a five year old computer and 300 €.

I think the bad thing about thouse generic perfomances results is that it has two flaws:
1) it does not take on to account TDP of CPU, But TDP means that CPU can withstand more heat befor it will have to downclock.
And here where Xeons CPUs shine over i7s.
While i7s can give u better generic test results in real world i7 can work on max performan only for short period while Xeons with good cooling ( Which new silinder Mac Pro has) can work on max performance 24/7 so in real world Xeos have better performance.
2) another thing is CPU cashe size. If ur cashe is full, for no real time task it will mean longer time of computations but for real time task like music production it will mean cracks and noises( even though that u CPU migh not be at full load).

[chilly7]

As I said, in single core performance several models are superior to the Mac Pro, though it seems surprising. However, it makes sense when you compare these new MacBooks to the last generation (Late 2013). Future benchmarks from other users will clarify. These MacBooks were just released.

According to that chart, the Macbook Pro @ 2.8 ghz is faster than the Mac Pro @ 3.7 Ghz for both single core and multi-core...

How can that be right?

I think one reason can be because of:
1) Turbo Boost
2) for a stupid reasons i7 is always one generation ahead of Xeons CPUs.

But because i7s has lower TDP they can not stay that long on max performance so in sort time Xeons will give better performance over i7s.

P.s.
In old days of ( Pentium 1,2,3 computers) i heared that new CPU needs time to work better. The same as new speakers need time to work better.
I do not know if it is true so i leave it at ur thoughts.
My new silender baseMac Pro got about 12.000 point under Geek bench 3 in 32bit test and my mother Mac Book Air 2012 13 inch base model got about 3.500 points.

[whyterabbyt]

The bad thing about thouse generic perfomances results is that it has two flaws:
1) it does not take on to account TDP of CPU, But TDP means that CPU can withstand more heat befor it will have to downclock.
And here where Xeons CPUs shine over i7s.
While i7s can give u better generic test results in real world i7 can work on max performan only for short period while Xeons with good cooling can work on max performance 24/7( Which new silinder Mac Pro has) so in real world Xeos have better performance.

Actually TDP is the heat generated, not how much it can stand.
But any decent benchmark results should reflect a processor's thermal behaviour. And since working with audio doesnt actually entail a 24/7 performance requirement its probably not that relevant to bring up.

Nice of you to compare 'an i7' with 'Xeons with good cooling' btw. Wouldnt want to make it fair and compare the latter to 'an i7 with good coooling,' would we.

) another thing is CPU cashe size. If ur cashe is full, for no real time task it will mean longer time of computations but for real time task like music production it will mean cracks and noises.

Well, no. A full cache doesnt mean cracks and noises. After a certain point in time, the cache is going to be full anyway. The issue is whether it contains the code you want (ie cache hit), and if not, how long it takes to access that (cache miss). But even cache misses, dont automatically mean cracks and noises, not if the audio buffers can be filled sufficiently fast. And cache misses are kind of inevitable anyway if you're running if, eg a lot of disparate code like, erm, multiple plugins.
Meanwhile, any decent benchmark will reflect a CPU's cache effectiveness.